Hi all-

A couple years ago a friend recommended that I read "The Company of One" by Paul Jarvis^[1]. We share a similar love of small, “do one thing well”-style businesses and he rightly saw the potential for some inspiration. While the title implies solopreneurs, I think this book actually taps into something larger — the power of small, focused teams and recognizing what's "enough". In our world of unicorns and multi-billion-dollar valuations, it's easy to forget that there are viable, lucrative paths to founding businesses that don’t require hyper-growth and hyper-stress.

One of the reasons I'm so excited about AI is its potential to enable an even wider canvas of business opportunities for solopreneurs, indie hackers, and small teams. My hypothesis is that AI's true power lies not in wholesale human replacement, but in collaboration, potentially enabling smaller teams to build products that would have required much larger organizations in the past. As I've pursued my own exploration of these new tech capabilities, I keep looking for the use cases that will unlock the creativity of everyday entrepreneurs.

To road test this hypothesis, I recently ran an AI-enhanced strategic planning process for my own business, using tools like Claude Projects, Perplexity, and Notion. In today’s article, I'll walk you through my experiment, share key learnings and insights, and explore the implications for the "company of one". Let's get to it.

One of my favorite parts of building a new product is mapping the user journey. It’s so simple, but it unlocks such a deep understanding of what needs to happen to get your user from their need to their solution. You can visualize the steps, rearrange, and think through the flow of their day. As I’ve dug into AI, I’ve found the user journey map to be an incredibly useful concept for thinking about best practices for integrating it into your work.

As I’ve said before, a common misconception of AI is that it’s an easy button — you ask it for something, and it gives you a perfect thing back. Visually that looks something like this:

But I find AI’s comparison to a hyper-capable, collaborative intern^[2] to be much more useful. You can delegate specific tasks to it, but ultimately the human is going to be responsible for some part of the project.

Much of my exploration of AI today is focused around trying to figure out what types of steps are best delegated — where are the places you get the most benefit, and where are the places that you’re better suited to take the wheel.

The Experiment: AI-Enhanced Strategic Planning

Strategic planning and roadmapping are fundamental to effective product leadership. Over the years, I've refined an approach that, while adaptable, always adheres to a few core principles. They're the backbone of how I think about planning:

• Outcome-oriented: The focus is always on the desired end result, not just the steps to get there. It's the difference between "not being hungry" (outcome) and "slicing bread" (output). Focusing solely on outputs can lead you astray from your true goals.
• Multiscale: I work backwards from long-term goals to determine near-term priorities. This approach provides crucial context for day-to-day decision-making. I typically start with a 3-5 year vision, then break it down into annual, quarterly, and even weekly milestones.
• Nested: I keep planning documents concise and navigable by creating separate, interconnected documents for different time scales. This structure allows for quick shifts between levels of detail without constant updates to higher-level concepts.

With these principles in mind, I set out to design an AI-enhanced strategic planning process. I had two main goals:

• create a cohesive plan for my own business
• build up Claude’s knowledge on my life so that it could more easily slot into ad hoc tasks without having to constantly re-share context

To do this, I leveraged three tools for different parts of the process:

• Claude Projects: Served as the primary workspace for drafting and refining planning documents.
• Notion: My “second brain” and core knowledge management space. Used for organizing influences and raw notes, and storing final versions of documents, providing a more navigable structure than Claude's artifact system.
• Perplexity: Employed for research, gathering ideas and references to inform the planning documents.

Setting the Stage

As a test I started by asking Claude how to approach the planning process. However, when left to its own devices I found that the AI's suggestions often included steps I knew weren't necessary for my approach. Once I shared my overall structure with Claude I found that the process moved much more quickly. This early hiccup highlighted an important lesson: while AI can be a great partner, at this stage of the tech’s development you often still need a human to orchestrate larger projects.

The AI-Enhanced Process

My first step was creating a project in Claude, setting up separate conversations for each artifact of my plan. After completing each artifact, I would add it to Notion and Claude's Project Knowledge — a crucial step for maintaining context across different planning levels. Here's how the process unfolded:

• 3-5 Year Vision: I started by working with Claude to draft long-term goals and vision. This document served as the foundation for all subsequent planning.
• Annual Milestones: With the long-term vision established, we moved on to setting annual milestones that would support the 3-5 year goals.
• Quarterly Plan: We then broke down the annual goals into a quarterly plan with monthly milestones.
• Monthly and Weekly Planning: Using the quarterly plan as a guide, I could then organize my months and weeks, maintaining alignment with the broader strategy.

The Collaborative Workflow

While working through this process, I found that each artifact conversation with Claude followed a consistent pattern:

• I'd begin by specifying the artifact I wanted to create and its structure, remind Claude of the context available in the project knowledge, then ask it what additional information it needed.
• Using Claude's questions as prompts, I'd provide an unstructured brain dump of thoughts and references.
• Claude would draft a synthesis, which we'd then iteratively refine with a focus on the strategic content. I’d often ask Claude to critique a given draft, asking for risks or holes in the plan. Armed with the critique, we’d then tighten up the strategy to mitigate the risks.
• For longer documents, I found it more manageable to work through this process section-by-section, as Claude's performance could decline when handling large amounts of content at once.
• Once the content was there, we’d move on to refining structure and tone until I had a final version I was happy with.

Using this process and workflow, I was able to work through a full planning cycle in about a week. While I don’t have a great apples-to-apples comparison for the time investment, anecdotally I found it was much easier to maintain momentum during the project. More often than not I felt like I was in a flow state, primarily due to the way I was able to leverage some of the key strengths of LLM-based AIs.

Key Learnings and Insights

After a week of experimenting with AI in my strategic planning process, I found a few takeaways that kept rising to the top. While I had some expectations going in, I was pleasantly surprised to uncover some new approaches that I’ve begun incorporating into my work.

Maintaining Creative Momentum

One of the first things I noticed was how helpful Claude was in keeping my creative juices flowing. We've all been there - staring at a blank page, not sure where to start. I found that in those moments, describing what I was trying to accomplish to Claude and asking what questions it had often broke through that mental block. These questions became prompts that enabled me to answer in a more unstructured manner, easing the burden. It was like having an always-available brainstorming partner, which could be a game-changer for solo entrepreneurs or small teams who don't have the luxury of bouncing ideas off colleagues whenever they want.

Data Synthesis and Theme Identification

A related, standout feature was Claude's ability to make sense of my jumbled thoughts and notes. I tend to start my planning process with a mess of ideas, scribbled notes, and half-formed concepts. Claude excelled at taking this chaos and organizing it into coherent themes. While I often had to massage those themes and remind it of nuances from my notes, these first passes saved me a ton of time that I usually spend just trying to organize my thoughts.

Rapid Iteration

The AI also proved to be a valuable tool for rapid iteration. In writing and planning, you often don't know if an idea works until you've fleshed it out a bit. Claude could quickly generate longer passages or more detailed plans, allowing me to assess the viability of ideas much faster than if I were writing everything out myself. For small businesses that need to stay nimble, this could be a huge advantage in adapting strategies quickly.

Critique and Refine

One of the most powerful techniques I leveraged was asking Claude to critique its own work. After we'd draft a strategy, I'd ask for feedback on it. This often revealed blind spots or potential risks I hadn't considered. It was like having a devil's advocate in the room, challenging assumptions and pushing for more robust planning.

The Acceleration Effect

As we progressed through the planning process (and our project context became more comprehensive), I noticed that Claude got better at understanding my goals and style. Its first drafts became more accurate and aligned with what I was aiming for, often needing minimal tweaking. While this required some upfront investment in providing context, it paid off in the later stages of planning and even in related tasks outside of the main planning process.

Limitations in Linear planning

That said, it wasn't all smooth sailing. I found that while Claude was great at high-level strategy and monthly planning, it struggled with more granular, day-to-day planning. When we tried to break things down to weekly or daily goals, the AI often lost track of context or forgot about other commitments I had mentioned. I found myself spending a lot of time re-explaining my schedule and constraints, to the point where it was often quicker to just do the detailed planning myself.

Nuance and Quality

This limitation highlights an important, related point: AI, at least in its current state, is best used as a collaborative tool rather than a full human replacement. Throughout the process, I often had to balance Claude's input with my own expertise and intuition. There were times when its suggestions, while logical, just didn't align with my deeper understanding of my business or personal goals. For pretty much every artifact we produced, I found myself having to pull key nuances through and ensure the overall quality of the final piece.

Learning to navigate this balance - knowing when to lean on AI's strengths and when to trust my own judgment - was perhaps the most valuable lesson from this experiment. As AI continues to evolve, I'm curious to see how it might improve at handling both more nuanced and more context-heavy tasks. My hunch is that, as it gains the ability to “see” more of our calendars and data, that it will be become much better at the “personal assistant” tasks like scheduling, but it will still require a human hand at the wheel for things like nuance and prioritization.

Implications for the "Company of One"

The integration of AI into strategic planning opens up exciting possibilities for solo entrepreneurs and small teams, potentially reshaping how we approach entrepreneurship in the AI age.

For the "Company of One," AI's ability to simulate collaborative thinking is transformative. It enables solo operators to refine their strategies and concepts with a depth previously challenging to achieve alone. Moreover, AI serves as an excellent starting point in areas where an entrepreneur might lack expertise, allowing a solopreneur to effectively "fill out" their team with AI-generated outputs in various domains. This could significantly reduce startup costs by delaying the need for early hires, enabling more effective bootstrapping.

However, harnessing these AI tools requires an important upfront investment. Spending time to help the AI understand your nuances, voice, and goals - the "soft" elements that make you and your business unique - is crucial. While this process might feel slow initially, it pays dividends as the AI gains a deeper understanding of your vision.

It's vital to approach AI-generated content with a critical eye. The biggest risk is the temptation to accept outputs wholesale. There's no substitute for your own taste, nuance, and expertise. In my experience, these tools excel at bringing projects from 0 to about 70-80% completion, freeing you up to focus on the crucial elements that truly make the output unique to your business and vision.

This dynamic creates an interesting shift in how solo entrepreneurs allocate their time and energy. Instead of starting from scratch in every aspect of your business, you can use AI to quickly generate solid foundations across various areas^[3]. Your role then becomes more about refinement, personalization, and strategic decision-making, allowing you to operate at a scale and with a level of sophistication previously associated with larger teams.

Looking ahead, as AI tools continue to evolve, we might see even more opportunities for solo entrepreneurs to compete effectively with larger organizations. The ability to rapidly iterate on ideas, get instant feedback, and explore multiple strategic options could level the playing field in unprecedented ways.

However, it's crucial to remember that AI is a tool, not a magic solution. The most successful "Companies of One" in the AI age will likely be those that strike a balance - leveraging AI's strengths while maintaining their unique human insights and creativity.

As we navigate this new frontier of AI-enhanced entrepreneurship, the key will be learning to dance with AI - knowing when to lead, when to follow, and always keeping sight of our own vision and values. For solo entrepreneurs and small teams willing to embrace this new paradigm, the potential for innovation, efficiency, and impact is enormous.

The journey of integrating AI into our strategic processes is just beginning. As these tools evolve, so too will our methods of collaboration with them. By staying curious, adaptable, and true to our unique perspectives, we can harness the power of AI to not just compete, but to redefine what's possible for the "Company of One" in the modern business landscape.